Re: [boost] Re: Live summary of regression tests.

From: Toon Knapen (toon.knapen_at_[hidden])
Date: 2003-02-10 09:01:05

On Saturday 08 February 2003 00:26, Rene Rivera wrote:
> >If the old run is no longer relevent (as run date October 17 suggests)
> >who would be responsible for removing it?
> 
> I would think the person who ran the test, or Beman, or David, or any admin,
> etc.

These are results from compilations by Markus Shoepflin on VisualAge version 5.
For a long time, he has not updated them anymore so let's ask him (he's in CC)

> >> I've been trying to figure out a good way to show that some test are
> newer
> >> than others, because sorting by date is not possible, and I like that
> >> suggestion very much :-]  I'd prefer a bit more granularity than you
> suggest
> >> though. How'bout this:
> >> 
> >>     GREEN -- tests ran in the last 48 hours.
> >>     YELLOW -- tests ran in the last 7 days.
> >>     RED -- tests ran in the last 14 days.
> >>     BLACK -- tests are older than 14 days.
> >
> >I tend to like default (black) text to indicate there is no error, and
> >only colour-code things that need attention.  In your scheme it is the
> >reverse, black being an 'abandoned' test suite.
> 
> Ah, sorry that was my mistake. I really meant to keep your colors with just
> the changed increments :-\  -- So it would be BLACK for 48 hours, and so on.

I think the traffic-light colors should suffice. I find adding black confusing.
Ideally the color green should only be used if the sources have not changed in the mean time
so in general the colors depend on the amount of change between the sourced use for the test
in respect to the sources in the CVS. But I can imagine that that's not possible or .... ?